Improvement of Iron Overload Parameters in Patients with Transfusion-Dependent β-Thalassemia Treated with Luspatercept: Data from the Phase 3b Long-Term Rollover Study Following the BELIEVE Trial

Introduction: Patients (pts) with transfusion-dependent (TD) β-thalassemia require lifelong red blood cell (RBC) transfusions, which can contribute to iron overload and increased morbidity and mortality. In the phase 3 BELIEVE trial (NCT02604433), treatment with luspatercept reduced transfusion burden (TB) with a manageable safety profile. At the end of the BELIEVE trial, pts receiving luspatercept were eligible for the open-label, phase 3b long-term follow-up (LTFU) study (NCT04064060). The aim of this analysis was to evaluate the effect of long-term luspatercept treatment on iron parameters in pts who entered the LTFU study from the BELIEVE trial.

Methods: This analysis included pts randomized to luspatercept during the BELIEVE trial (luspatercept group) and pts randomized to placebo who crossed over to receive luspatercept after unblinding (crossover group). For crossover pts, treatment timing was defined from initiation of luspatercept treatment. Serum ferritin (SF) levels, liver iron concentration (LIC) by MRI (derived from T2*, R2*, or R2 method), and changes in iron chelation therapy (ICT) use were evaluated per 48-wk intervals alongside changes in RBC TB. Mean changes from baseline (BL) were calculated using values from evaluable pts. BL values for SF and ICT use were calculated for the 12 wk prior to first dose and at the last assessment on/before the first dose for LIC; BL values for crossover pts were prior to starting placebo. The data cutoff was Jan 28, 2024.

Results: Median (range) treatment duration over the combined BELIEVE and LTFU studies was 229.1 wk (1.7–388.6) for the luspatercept group and 225.9 wk (6.1–286.6) for the crossover group. Median (range) follow-up was 303.3 wk (1.7–392.0) and 268.1 wk (11.1–286.6), respectively. At data cutoff, 111 pts had ongoing treatment: 72/224 (32.1%) in the luspatercept group and 39/92 (42.4%) in the crossover group. Concomitant ICT was received by > 98% of pts, with 72% (n = 161) in the luspatercept group and 74% (n = 68) in the crossover group receiving deferasirox.

In the luspatercept group, mean change from BL in SF ranged from −247.2 μg/L to −541.3 μg/L during wk 48–384. Overall mean change from BL in SF at efficacy cutoff was −374.1 μg/L (standard deviation [SD] 1209.4) from a median BL of 1416.0 μg/L (range, 88.0–6400.0). In the crossover group, SF levels began to decrease from wk 144; however, overall SF was similar to BL: overall mean change from BL at efficacy cutoff was 21.2 μg/L (SD 1128.8) from a median BL of 1241.0 μg/L (range, 136.0–6400.0). In the luspatercept group, 42.6% of pts with BL SF ≥ 1000 μg/L shifted to SF < 1000 μg/L by efficacy cutoff. Similarly, 34.9% of pts in the crossover group shifted to < 1000 μg/L versus BL.

Median (range) LIC at BL was 5.9 mg/g dry weight (dw) (0.8–42.0) in the luspatercept group and 4.6 mg/g dw (0.2–43.0) in the crossover group. LIC values remained near BL up to wk 96 in the luspatercept group, after which only small reductions from BL were observed (mean change ranged from −3.1 to 0.2 mg/g dw during wk 96–336). In the crossover group, LIC values were similar to BL; mean change from BL during wk 96–288 ranged from −0.2 to 1.5 mg/g dw.

ICT use decreased from BL in the luspatercept group, particularly for deferasirox, for which the mean change from BL in daily dose ranged from −103.4 mg to −192.4 mg during wk 144–240. In the crossover group, mean daily deferasirox dose decreased through wk 288, with the highest reductions from BL seen at wk 48 (−252.5 mg), wk 96 (−204.9 mg), and wk 240 (−198.0 mg).

A decrease in RBC TB from BL was maintained in both the luspatercept (through wk 336) and crossover (through wk 240) groups. At each interval, the mean reduction in the number of RBC units transfused ranged from 4.8 to 7.8 units/48 wk and 5.1 to 6.9 units/48 wk for the luspatercept and crossover groups, respectively, representing a mean reduction from BL of ~20% (mean reductions ranging from 17.4% to 26.8% and 17.0% to 25.0%, respectively).

Conclusions: Long-term treatment with luspatercept (median duration > 4 y) resulted in decreased SF maintained below BL levels. Furthermore, LIC remained stable despite lower daily doses of deferasirox. The stabilization or improvement in iron parameters is likely attributable to the sustained reduction in RBC TB with luspatercept treatment. This analysis supports the overall long-term benefit of luspatercept in pts with TD β‑thalassemia, potentially alleviating iron overload and preventing future complications.